Background: To compare the prognostic value of estimated glomerular filtration rate, cystatin-C, an alternative renal biomarker, and their combination, in an outpatient population with heart failure.Estimated glomerular filtration rate is routinely used to assess renal function in heart failure patients. We recently demonstrated that the Cockroft-Gault formula is the best among the most commonly used estimated glomerular filtration rate formulas for predicting heart failure prognosis. Methodology/Principal Findings: A total of 879 consecutive patients (72% men, age 70.4 years [P25-75 60.5-77.2]) were studied. The etiology of heart failure was mainly ischemic heart disease (52.7%). The left ventricular ejection fraction was 34% (P25-75 26-43%). Most patients were New York Heart Association class II (65.8%) or III (25.9%). During a median follow-up of 3.46 years (P25-75 1.85-5.05), 312 deaths were recorded. In an adjusted model, estimated glomerular filtration rate and cystatin-C showed similar prognostic value according to the area under the curve (0.763 and 0.765, respectively). In Cox regression, the multivariable analysis hazard ratios were 0.99 (95% CI: 0.98-1, P = 0.006) and 1.14 (95% CI: 1.02-1.28, P = 0.02) for estimated glomerular filtration rate and cystatin-C, respectively. Reclassification, assessed by the integration discrimination improvement and the net reclassification improvement indices, was poorer with cystatin-C (-0.5 [-1.0;-0.1], P = 0.024 and -4.9 [-8.8;-1.0], P = 0.013, respectively). The value of cystatin-C over estimated glomerular filtration rate for risk-stratification only emerged in patients with moderate renal dysfunction (eGFR 30-60 ml/min/1.73 m2, chi-square 12.9, P<0.001). Conclusions/Significance: Taken together, the results indicate that estimated glomerular filtration rate and cystatin-C have similar long-term predictive values in a real-life ambulatory heart failure population. Cystatin-C seems to offer improved prognostication in heart failure patients with moderate renal dysfunction. © 2012 Zamora et al.
|Publication status||Published - 11 Dec 2012|